Reliable detection and continuous monitoring of memory dysfunction in mild cognitive impairment and healthy aging through adaptive computational phenotyping
5 affiliations
- 1Department of Psychology, University of Washington, Seattle, Washington, United States of America
- 2Graduate Program in Neuroscience, University of Washington, Seattle, Washington, United States of America
- 3Precision Cognition Labs, Groningen, The Netherlands
- 4Departments of Radiology and Neurology, University of Washington, Seattle, Washington, United States of America
- 5Department of Experimental Psychology, University of Groningen, Groningen, The Netherlands
Abstract
With the rising prevalence of age-related memory impairments, efficiently detecting and monitoring decline is increasingly urgent. Unfortunately, traditional assessment methods fall short of these needs, as they typically require in-person administration and cannot be repeated frequently. Here, we demonstrate that remote identification and monitoring of abnormal memory function is possible by combining an online assessment platform with computational phenotyping, allowing repeatable, unsupervised remote observations from patients. Fifty-one well-characterized older individuals, including 24 patients with amnestic mild cognitive impairment and 27 age- and education-matched healthy controls, completed a series of longitudinal, unsupervised, remote weekly 8-minute online memory assessments for up to one year. Weekly test data were fit to a formal model of memory consolidation and forgetting, yielding an individualized index of memory function, the Seattle-Groningen Memory Assessment (SGMA) score. The SGMA score was found to be reliable, with a mean correlation of r = 0.70 across assessments. The score was also found to be stable across different study materials, and only barely affected by practice effects, which averaged to a 0.2% increase per assessment. Finally, the SGMA score was found to be diagnostic, being capable of detecting mild cognitive impairment with up to 87% accuracy. These findings show that model-based, adaptive assessments can support high-frequency, remote detection and scalable longitudinal monitoring of early memory decline, providing a new way to assess memory decline trajectories in healthy aging and dementia. Author summary: Memory decline is a central feature of aging and an early marker of neurodegenerative disease, yet current methods for assessing memory are poorly suited for frequent, long-term monitoring. Standard neuropsychological tests are time-consuming, administered in person, and cannot be repeated often without substantial practice effects. As a result, changes in memory function are often detected too late and their trajectories cannot be traced. In this study, we show that memory function can be reliably assessed and monitored remotely using a brief, unsupervised online task combined with computational modeling. Older adults, including individuals with mild cognitive impairment, completed short weekly memory assessments from home over periods of up to one year. Rather than relying on raw test scores, we used a digital twin approach: each participant’s data was fitted to a mathematical model of memory consolidation, and its forgetting parameter was used to derive a precise measure of memory function, the Seattle–Groningen Memory Assessment (SGMA) score. We find that this model-based score is reliable over time (r = 0.70), largely unaffected by repeated testing, and able to distinguish individuals with mild cognitive impairment from healthy older adults with high accuracy (up to 87%). These results demonstrate that digital, model-driven cognitive assessments can support scalable, high-frequency monitoring of memory decline. Such approaches may complement existing clinical tools and enable earlier detection and more precise tracking of cognitive changes in aging and dementia.
Related
- Interpretable computational SHG microscopy enables quantitative tissue phenotyping of cardiac remodeling in nonischemic heart failure
- Longitudinal computational phenotyping reveals a stable marker for depression symptoms
- The capability of very high-resolution satellite imagery for plot-level early wheat stem rust disease detection, monitoring, and phenotyping in Ethiopia
- Premature Ventricular Contractions at Extreme Altitude: Continuous Electrocardiographic Monitoring on Mount Everest.
- Continuous monitoring of deficit irrigation in avocado across two contrasting rainfall years using sensor networks, telemetry, and machine learning
BibTeX
@article{hake2026reliable,
title = {{Reliable detection and continuous monitoring of memory dysfunction in mild cognitive impairment and healthy aging through adaptive computational phenotyping}},
author = {Holly S. Hake and Maarten van der Velde and Thomas J. Grabowski and Hedderik van Rijn and Andrea Stocco},
journal = {PLOS Digital Health},
year = {2026},
month = sep,
doi = {10.1371/journal.pdig.0001686},
url = {https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001686},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Reliable detection and continuous monitoring of memory dysfunction in mild cognitive impairment and healthy aging through adaptive computational phenotyping
AU - Holly S. Hake
AU - Maarten van der Velde
AU - Thomas J. Grabowski
AU - Hedderik van Rijn
AU - Andrea Stocco
PY - 2026
DA - 2026/09/15
JO - PLOS Digital Health
DO - 10.1371/journal.pdig.0001686
UR - https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001686
AB - With the rising prevalence of age-related memory impairments, efficiently detecting and monitoring decline is increasingly urgent. Unfortunately, traditional assessment methods fall short of these needs, as they typically require in-person administration and cannot be repeated frequently. Here, we demonstrate that remote identification and monitoring of abnormal memory function is possible by combining an online assessment platform with computational phenotyping, allowing repeatable, unsupervised remote observations from patients. Fifty-one well-characterized older individuals, including 24 patients with amnestic mild cognitive impairment and 27 age- and education-matched healthy controls, completed a series of longitudinal, unsupervised, remote weekly 8-minute online memory assessments for up to one year. Weekly test data were fit to a formal model of memory consolidation and forgetting, yielding an individualized index of memory function, the Seattle-Groningen Memory Assessment (SGMA) score. The SGMA score was found to be reliable, with a mean correlation of r = 0.70 across assessments. The score was also found to be stable across different study materials, and only barely affected by practice effects, which averaged to a 0.2% increase per assessment. Finally, the SGMA score was found to be diagnostic, being capable of detecting mild cognitive impairment with up to 87% accuracy. These findings show that model-based, adaptive assessments can support high-frequency, remote detection and scalable longitudinal monitoring of early memory decline, providing a new way to assess memory decline trajectories in healthy aging and dementia. Author summary: Memory decline is a central feature of aging and an early marker of neurodegenerative disease, yet current methods for assessing memory are poorly suited for frequent, long-term monitoring. Standard neuropsychological tests are time-consuming, administered in person, and cannot be repeated often without substantial practice effects. As a result, changes in memory function are often detected too late and their trajectories cannot be traced. In this study, we show that memory function can be reliably assessed and monitored remotely using a brief, unsupervised online task combined with computational modeling. Older adults, including individuals with mild cognitive impairment, completed short weekly memory assessments from home over periods of up to one year. Rather than relying on raw test scores, we used a digital twin approach: each participant’s data was fitted to a mathematical model of memory consolidation, and its forgetting parameter was used to derive a precise measure of memory function, the Seattle–Groningen Memory Assessment (SGMA) score. We find that this model-based score is reliable over time (r = 0.70), largely unaffected by repeated testing, and able to distinguish individuals with mild cognitive impairment from healthy older adults with high accuracy (up to 87%). These results demonstrate that digital, model-driven cognitive assessments can support scalable, high-frequency monitoring of memory decline. Such approaches may complement existing clinical tools and enable earlier detection and more precise tracking of cognitive changes in aging and dementia.
SN - 2767-3170
ER -